[2026-01-06 06:31:33,733] [DEBUG] [axolotl.utils.config.log_gpu_memory_usage:127] [PID:5347] baseline 0.000GB () [2026-01-06 06:31:33,734] [INFO] [axolotl.cli.config.load_cfg:248] [PID:5347] config: { "activation_offloading": false, "axolotl_config_path": "sft-axolotl-olmo3-7b-think.yaml", "base_model": "allenai/Olmo-3-1025-7B", "base_model_config": "allenai/Olmo-3-1025-7B", "batch_size": 8, "bf16": true, "capabilities": { "bf16": true, "compute_capability": "sm_90", "fp8": false, "n_gpu": 2, "n_node": 1 }, "chat_template": "jinja", "chat_template_jinja": "/workspace/data/model-output/chat_template.jinja", "context_parallel_size": 2, "dataloader_num_workers": 2, "dataloader_pin_memory": true, "dataloader_prefetch_factor": 256, "dataset_prepared_path": "last_run_prepared", "dataset_processes": 48, "datasets": [ { "chat_template": "tokenizer_default", "field_messages": "messages", "message_property_mappings": { "content": "content", "role": "role" }, "path": "dataset-tfs-mk-IMP-SOS-processed-olmo3-think.jsonl", "roles": { "assistant": [ "assistant" ], "system": [ "system" ], "user": [ "user" ] }, "roles_to_train": [ "assistant" ], "trust_remote_code": false, "type": "chat_template" } ], "ddp": true, "device": "cuda:0", "device_map": { "": 0 }, "dion_rank_fraction": 1.0, "dion_rank_multiple_of": 1, "env_capabilities": { "torch_version": "2.7.1" }, "eval_batch_size": 1, "eval_causal_lm_metrics": [ "sacrebleu", "comet", "ter", "chrf" ], "eval_max_new_tokens": 128, "eval_sample_packing": true, "eval_steps": 0.5, "eval_table_size": 0, "evals_per_epoch": 1, "experimental_skip_move_to_device": true, "flash_attention": true, "fp16": false, "gradient_accumulation_steps": 4, "gradient_checkpointing": true, "gradient_checkpointing_kwargs": { "use_reentrant": true }, "group_by_length": false, "hub_model_id": "Auditt/O37BB", "include_tkps": true, "learning_rate": 1e-05, "lisa_layers_attribute": "model.layers", "load_best_model_at_end": false, "load_in_4bit": false, "load_in_8bit": false, "local_rank": 0, "logging_steps": 1, "lora_dropout": 0.0, "loraplus_lr_embedding": 1e-06, "lr_scheduler": "cosine", "mean_resizing_embeddings": false, "micro_batch_size": 1, "model_config_type": "olmo3", "num_epochs": 2.0, "optimizer": "adamw_torch", "output_dir": "/workspace/data/model-output-base", "pad_to_sequence_len": true, "pretrain_multipack_attn": true, "profiler_steps_start": 0, "qlora_sharded_model_loading": false, "ray_num_workers": 1, "resources_per_worker": { "GPU": 1 }, "ring_attn_func": "varlen_llama3", "sample_packing": true, "sample_packing_bin_size": 200, "sample_packing_group_size": 100000, "save_only_model": false, "save_safetensors": true, "sequence_len": 60000, "shuffle_before_merging_datasets": false, "shuffle_merged_datasets": true, "skip_prepare_dataset": false, "streaming_multipack_buffer_size": 10000, "strict": false, "tensor_parallel_size": 1, "tf32": true, "tiled_mlp_use_original_mlp": true, "tokenizer_config": "allenai/Olmo-3-1025-7B", "tokenizer_save_jinja_files": true, "tokens": [ "\ud801\udd32", "\ud801\udd3e", "\u3009", "\ud835\udf0e", "\u22c1", "\ud801\udd60", "\ud801\udd5c", "\ud801\udd38", "\u2227", "\u2265", "\ud801\udd5f", "\ud801\udd56", "\u27c2", "\ud801\udd4f", "\u22c0", "\ud801\udd63", "\ud801\udd43", "\ud801\udd59", "\ud801\udd55", "\u03c7", "\ud801\udd4a", "\u3008", "\ud801\udd50", "\ud801\udd3b", "\ud801\udd40", "\ud801\udd33", "\u2260", "\ud801\udd37", "\u2264", "\ud801\udd5e", "\ud801\udd31", "\ud801\udd42", "\u21a6", "\ud801\udd4e", "\u2192", "\ud801\udd5b", "\ud801\udd30", "\u03b5" ], "torch_dtype": "torch.bfloat16", "train_on_inputs": false, "trl": { "log_completions": false, "mask_truncated_completions": false, "ref_model_mixup_alpha": 0.9, "ref_model_sync_steps": 64, "scale_rewards": true, "sync_ref_model": false, "use_vllm": false, "vllm_server_host": "0.0.0.0", "vllm_server_port": 8000 }, "trust_remote_code": true, "use_ray": false, "val_set_size": 0.1, "vllm": { "device": "auto", "dtype": "auto", "gpu_memory_utilization": 0.9, "host": "0.0.0.0", "port": 8000 }, "weight_decay": 0.0, "world_size": 2 } [2026-01-06 06:31:34,116] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:278] [PID:5347] EOS: 100257 / <|endoftext|> [2026-01-06 06:31:34,116] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:279] [PID:5347] BOS: None / None [2026-01-06 06:31:34,117] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:280] [PID:5347] PAD: 100277 / <|pad|> [2026-01-06 06:31:34,117] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:281] [PID:5347] UNK: 100257 / <|endoftext|> [2026-01-06 06:31:48,563] [INFO] [axolotl.utils.data.shared.load_preprocessed_dataset:470] [PID:5347] Loading prepared dataset from disk at last_run_prepared/521442581534a9837f30b55bdde4d057... [2026-01-06 06:31:48,578] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:404] [PID:5347] total_num_tokens: 4_377_664 [2026-01-06 06:31:48,586] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:422] [PID:5347] `total_supervised_tokens: 3_345_873` [2026-01-06 06:31:49,583] [DEBUG] [axolotl.utils.samplers.multipack.__len__:458] [PID:5347] generate_batches time: 0.4595468044281006 [2026-01-06 06:31:50,030] [DEBUG] [axolotl.utils.samplers.multipack.__len__:458] [PID:5347] generate_batches time: 0.44755101203918457 [2026-01-06 06:31:50,477] [DEBUG] [axolotl.utils.samplers.multipack.__len__:458] [PID:5347] generate_batches time: 0.44657182693481445 [2026-01-06 06:31:50,952] [DEBUG] [axolotl.utils.samplers.multipack.__len__:458] [PID:5347] generate_batches time: 0.4744563102722168 [2026-01-06 06:31:52,230] [INFO] [axolotl.utils.samplers.multipack.calc_min_len:434] [PID:5347] gather_len_batches: [78, 78] [2026-01-06 06:31:52,231] [WARNING] [py.warnings._showwarnmsg:110] [PID:5347] /root/miniconda3/envs/py3.11/lib/python3.11/site-packages/torch/distributed/distributed_c10d.py:4631: UserWarning: No device id is provided via `init_process_group` or `barrier `. Using the current device set by the user. warnings.warn( # warn only once [2026-01-06 06:31:52,413] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:481] [PID:5347] data_loader_len: 9 [2026-01-06 06:31:52,429] [INFO] [axolotl.utils.trainer.calc_sample_packing_eff_est:497] [PID:5347] sample_packing_eff_est across ranks: [0.9235578179359436, 0.9235578179359436] [2026-01-06 06:31:52,430] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:509] [PID:5347] sample_packing_eff_est: None [2026-01-06 06:31:52,430] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:520] [PID:5347] total_num_steps: 36 [2026-01-06 06:31:52,542] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:404] [PID:5347] total_num_tokens: 39_860_654 [2026-01-06 06:31:53,267] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:422] [PID:5347] `total_supervised_tokens: 30_599_793` [2026-01-06 06:31:54,333] [DEBUG] [axolotl.utils.samplers.multipack.__len__:458] [PID:5347] generate_batches time: 0.5050613880157471 [2026-01-06 06:31:54,830] [DEBUG] [axolotl.utils.samplers.multipack.__len__:458] [PID:5347] generate_batches time: 0.49697208404541016 [2026-01-06 06:31:55,317] [DEBUG] [axolotl.utils.samplers.multipack.__len__:458] [PID:5347] generate_batches time: 0.48656487464904785 [2026-01-06 06:31:55,795] [DEBUG] [axolotl.utils.samplers.multipack.__len__:458] [PID:5347] generate_batches time: 0.47708654403686523 [2026-01-06 06:31:55,796] [INFO] [axolotl.utils.samplers.multipack.calc_min_len:434] [PID:5347] gather_len_batches: [699, 697] [2026-01-06 06:31:55,797] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:481] [PID:5347] data_loader_len: 87 [2026-01-06 06:31:55,798] [INFO] [axolotl.utils.trainer.calc_sample_packing_eff_est:497] [PID:5347] sample_packing_eff_est across ranks: [0.9517825841903687, 0.947709321975708] [2026-01-06 06:31:55,798] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:509] [PID:5347] sample_packing_eff_est: 0.96 [2026-01-06 06:31:55,802] [DEBUG] [axolotl.utils.trainer.calculate_total_num_steps:520] [PID:5347] total_num_steps: 348 [2026-01-06 06:31:55,802] [INFO] [axolotl.utils.data.sft._prepare_standard_dataset:121] [PID:5347] Maximum number of steps set at 348 [2026-01-06 06:31:55,821] [DEBUG] [axolotl.train.setup_model_and_tokenizer:65] [PID:5347] Loading tokenizer... allenai/Olmo-3-1025-7B [2026-01-06 06:31:56,142] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:278] [PID:5347] EOS: 100257 / <|endoftext|> [2026-01-06 06:31:56,142] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:279] [PID:5347] BOS: None / None [2026-01-06 06:31:56,142] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:280] [PID:5347] PAD: 100277 / <|pad|> [2026-01-06 06:31:56,142] [DEBUG] [axolotl.loaders.tokenizer.load_tokenizer:281] [PID:5347] UNK: 100257 / <|endoftext|> [2026-01-06 06:31:56,144] [DEBUG] [axolotl.train.setup_model_and_tokenizer:74] [PID:5347] Loading model [2026-01-06 06:31:56,194] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_evaluation_loop:87] [PID:5347] Patched Trainer.evaluation_loop with nanmean loss calculation [2026-01-06 06:31:56,195] [DEBUG] [axolotl.monkeypatch.transformers.trainer_loss_calc.patch_maybe_log_save_evaluate:138] [PID:5347] Patched Trainer._maybe_log_save_evaluate with nanmean loss calculation [2026-01-06 06:31:56,197] [DEBUG] [axolotl.monkeypatch.transformers.trainer_context_parallel.patch_prepare_context_parallel_inputs:66] [PID:5347] Patched Trainer._prepare_context_parallel_inputs for FlashAttention + CP [2026-01-06 06:31:56,198] [INFO] [axolotl.loaders.patch_manager._apply_multipack_patches:301] [PID:5347] Applying multipack dataloader patch for sample packing... Loading checkpoint shards: 0%| | 0/3 [00:00